CN103603776B - A kind of prediction algorithm of wind generating set pitch control Security - Google Patents
A kind of prediction algorithm of wind generating set pitch control Security Download PDFInfo
- Publication number
- CN103603776B CN103603776B CN201310603969.3A CN201310603969A CN103603776B CN 103603776 B CN103603776 B CN 103603776B CN 201310603969 A CN201310603969 A CN 201310603969A CN 103603776 B CN103603776 B CN 103603776B
- Authority
- CN
- China
- Prior art keywords
- layer
- blade
- neural network
- input
- hidden
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- Y02E10/722—
Landscapes
- Wind Motors (AREA)
Abstract
A prediction algorithm for wind generating set pitch control Security, adopt the possibility of BP three-layer neural network to the attached ice of blade to predict, described BP three-layer neural network comprises input layer, hidden layer and output layer; Input layer comprises 8 input nodes, is respectively: wind speed x1, air temperature x2, humidity of the air x3, change propeller angle x4, wheel speed x5, the spatial position x6 of blade, fan vibration numerical value x7, blade material coefficient x8; Hidden layer includes 3 node z1-z3; Output layer comprises 1 output node: icing rate v; The present invention extends unit lodicule working life, improves unit generation efficiency.
Description
Technical field
The invention belongs to technical field of wind power, be specifically related to a kind of prediction algorithm of wind generating set pitch control Security.
Background technique
Wind energy is as a kind of clean renewable energy sources, non-renewable energy resources environmental pollution problem can be solved to a great extent, especially exhaustion is faced in fossil energy, greenhouse gas emission day by day increases, when having had a strong impact on global climate, the application and development of wind technology has extremely been paid close attention in countries in the world.There are the grass resources of the length and breadth of land and very long shore line in China, and wind energy content enriches, but to determine the physical environment of wind field general all relatively more severe due to the feature of wind energy resources self, the more north being distributed in severe cold and moist coastal region.
When wind power generating set is run as humid air, rainfall or ice and snow weather at low ambient temperatures, freezing phenomena will be there is.The attached ice of wind power generating set blade not only can produce ice and carry, and can affect the life-span of blade.If blade band ice runs, more very large harm can be produced to unit: (1), if cause unit directly to be shut down because blade has ice, will make the unit generation amount being in low temperature area for a long time greatly reduce; (2) after the attached ice of blade, because the attached ice thickness in each cross section of blade differs, can directly affect the load of wind power generating set and exert oneself, make the decrease of power generation of unit; (3) the attached ice of blade, not only can produce potential safety hazard to blower fan self, also can threaten to resources such as field personnel, local resident and livestocks.
Summary of the invention
For solving the problem, the invention provides a kind of prediction algorithm of wind generating set pitch control Security, being intended to extend unit lodicule working life, improving unit generation efficiency.
Mentality of designing of the present invention is: the present invention to utilize wind speed, wheel speed, blade in spatial position, becomes the collection of the series of parameters such as propeller angle, air temperature, humidity of the air, fan vibration numerical value and blade material coefficient, predicts the attached ice possibility of blade by adopting BP three-layer neural network technology.
Technological scheme of the present invention is specific as follows:
A prediction algorithm for wind generating set pitch control Security, adopt the possibility of BP three-layer neural network to the attached ice of blade to predict, described BP three-layer neural network comprises input layer, hidden layer and output layer; Input layer comprises 8 input nodes, is respectively: wind speed x1, air temperature x2, humidity of the air x3, change propeller angle x4, wheel speed x5, the spatial position x6 of blade, fan vibration numerical value x7, blade material coefficient x8; Hidden layer includes 3 node z1-z3; Output layer comprises 1 output node: icing rate v; Input node, the function relation of hidden node and output node is as follows:
X
irepresent the integrated value that may affect icing data of unit collection, f1, f2 are the intrinsic parameter of neuron network, and n is constant, and i is a certain parameter, and the weights between input layer and hidden layer are h, and the weights of hidden layer and output layer are w, z
kfor hidden node, v is icing rate, and k is three blade numberings; Weight w, f1, the f2 of weights h, hidden layer and output layer between described input layer and hidden layer are drawn by software emulation;
The job step of described BP three-layer neural network is:
First BP three-layer neural network is tested, after the performance of BP three-layer neural network and error all converge to certain standard, utilize the BP three-layer neural network tested to carry out attached ice prediction; Then using new ice-formation condition as input, every sheet blade freezes thickness to utilize BP three-layer neural network to predict; Finally when v value exceedes limit value vmax certain hour t
1after, can think that fan blade surface freezes, the thickness frozen runs up to certain hour t
2after, unit carries out the deicing action of being correlated with.Described deicing action can select any known mode to carry out.But preferably use the change oar Self-Protection Subsystem in embodiment to carry out deicing.
Beneficial effect of the present invention is: a kind of prediction algorithm of wind generating set pitch control Security, solves wind power generating set when running under rugged environment, the problem such as short, unit generation efficiency reduction in the blade working life caused; Ensure that the safety of field personnel, local resident and livestock; Improve the safety and reliability of wind power generating set.
Accompanying drawing explanation
The present invention has accompanying drawing 4 width.
Fig. 1 is BP three-layer neural network topological diagram;
Fig. 2 is BP three-layer neural network workflow diagram;
Fig. 3 becomes oar Self-Protection Subsystem control logic figure in embodiment;
Fig. 4 becomes oar Self-Protection Subsystem figure in embodiment.
Embodiment
Below in conjunction with the present embodiment, the present invention is further described:
A prediction algorithm for wind generating set pitch control Security, adopt the possibility of BP three-layer neural network to the attached ice of blade to predict, described BP three-layer neural network comprises input layer, hidden layer and output layer; Input layer comprises 8 input nodes, is respectively: wind speed x1, air temperature x2, humidity of the air x3, change propeller angle x4, wheel speed x5, the spatial position x6 of blade, fan vibration numerical value x7, blade material coefficient x8; Hidden layer includes 3 node z1-z3; Output layer comprises 1 output node: icing rate v; Input node, the function relation of hidden node and output node is as follows:
X
irepresent the integrated value that may affect icing data of unit collection, f1, f2 are the intrinsic parameter of neuron network, and n is constant, and i is a certain parameter, and the weights between input layer and hidden layer are h, and the weights of hidden layer and output layer are w, z
kfor hidden node, v is icing rate, and k is three blade numberings; Weight w, f1, the f2 of weights h, hidden layer and output layer between described input layer and hidden layer are drawn by software emulation;
The job step of described BP three-layer neural network is:
First BP three-layer neural network is tested, after the performance of BP three-layer neural network and error all converge to certain standard, utilize the BP three-layer neural network tested to carry out attached ice prediction; Then using new ice-formation condition as input, every sheet blade freezes thickness to utilize BP three-layer neural network to predict; Finally when v value exceedes limit value vmax certain hour t
1after, can think that fan blade surface freezes, the thickness frozen runs up to certain hour t
2after, unit carries out the deicing action of being correlated with.Described deicing action can select any known mode to carry out.But preferably use change oar Self-Protection Subsystem to carry out deicing.
Described change oar Self-Protection Subsystem, comprising: be with the main control PLC of the logic that opens ice, the change oar frequency variator being with the prediction algorithm that opens ice, electric current device, drive the motor of blade rotation; Described band the open ice input end of change oar frequency variator of prediction algorithm of main control PLC and the band of logic that opens ice is connected, the open ice output terminal of change oar frequency variator of prediction algorithm of band is connected with driving the motor of blade rotation, described current transformer and be with the change oar frequency variator of the prediction algorithm that opens ice to be connected.
The specific works step of this system is:
According to the prediction algorithm becoming oar Security, when the blade thickness that freezes reaches the scope of automatic de-icing, or when automatic de-icing arrives cycle time, system enters initiatively deicing step;
Described active deicing step comprises:
(1) current transformer in blower fan detects blower fan and whether is in generator operation state, if now blower fan is in low wind outage state, then by being with the change oar frequency variator of the prediction algorithm that opens ice to make the motor of drive blade rotation accelerate to the racing speed of permission;
(2) if the current transformer detection blower fan in blower fan is in running order, then drive the motor of blade rotation to enter emergency feathering state by the open ice change oar Frequency Converter Control of prediction algorithm of band, the open ice main control PLC of logic of band records the vibration values in cabin in feathering process by the open ice change oar frequency variator of prediction algorithm of control cincture;
(3) after feathering completes, if the vibration values of record is less than the critical value of regulation, then think that this feathering deicing is invalid, repeat (1) (2) process, if be still less than the critical value of regulation in triplicate, then alerts triggered, directly enters initiatively collision block link;
Described active collision block link specifically comprises:
The open ice main control PLC of logic of band switches to speed control mode by the open ice change oar frequency variator of prediction algorithm of control cincture by driving the motor of blade rotation, becomes oar with increasing velocity;
After blade move angle is greater than 6 degree, the open ice main control PLC of logic of band drives the motor of blade rotation with 0.3deg/s speed impacts collision block by the open ice change oar Frequency Converter Control of prediction algorithm of control cincture, when blade rotational speed be less than 0.05deg/s and torque be greater than 50Nm continue 3s time, think that blade has arrived collision block stop position, stop becoming oar;
(4) remove related data and remove warning, terminating deicing.
Claims (1)
1. a prediction algorithm for wind generating set pitch control Security, is characterized in that: adopt the possibility of BP three-layer neural network to the attached ice of blade to predict, described BP three-layer neural network comprises input layer, hidden layer and output layer; Input layer comprises 8 input nodes, is respectively: wind speed x1, air temperature x2, humidity of the air x3, change propeller angle x4, wheel speed x5, the spatial position x6 of blade, fan vibration numerical value x7, blade material coefficient x8; Hidden layer includes 3 node z1-z3; Output layer comprises 1 output node: icing rate v; Input node, the function relation of hidden node and output node is as follows:
X
irepresent the integrated value that may affect icing data of unit collection, f1, f2 are the intrinsic parameter of neuron network, and n is constant, and i is a certain parameter, and the weights between input layer and hidden layer are h, and the weights of hidden layer and output layer are w, z
kfor hidden node, v is icing rate, and k is three blade numberings; Weights h between described input layer and hidden layer and the weight w of hidden layer and output layer are drawn by software emulation;
The job step of described BP three-layer neural network is:
First BP three-layer neural network is tested, after the performance of BP three-layer neural network and error all converge to certain standard, utilize the BP three-layer neural network tested to carry out attached ice prediction; Then using new ice-formation condition as input, every sheet blade freezes thickness to utilize BP three-layer neural network to predict; Finally when v value exceedes limit value vmax certain hour t
1after, can think that fan blade surface freezes, freeze through certain hour t
2after, unit carries out the deicing action of being correlated with.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310603969.3A CN103603776B (en) | 2013-11-23 | 2013-11-23 | A kind of prediction algorithm of wind generating set pitch control Security |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310603969.3A CN103603776B (en) | 2013-11-23 | 2013-11-23 | A kind of prediction algorithm of wind generating set pitch control Security |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103603776A CN103603776A (en) | 2014-02-26 |
| CN103603776B true CN103603776B (en) | 2016-03-02 |
Family
ID=50122024
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310603969.3A Active CN103603776B (en) | 2013-11-23 | 2013-11-23 | A kind of prediction algorithm of wind generating set pitch control Security |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103603776B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019086287A1 (en) * | 2017-10-30 | 2019-05-09 | fos4X GmbH | Method for forecasting the yield of a wind farm under icing conditions |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104005917B (en) * | 2014-04-30 | 2016-12-07 | 叶翔 | Method and system fan condition being predicted based on Bayesian inference mode |
| CN110147811A (en) * | 2019-04-02 | 2019-08-20 | 宜通世纪物联网研究院(广州)有限公司 | Fan blade prediction technique and system based on time window mixed model |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101886617A (en) * | 2010-06-07 | 2010-11-17 | 三一电气有限责任公司 | Wind generating set and blade deicing system thereof |
| CN102003353A (en) * | 2010-12-10 | 2011-04-06 | 重庆大学 | Deicing method for blades of large-scale wind driven generator |
| CN102622482A (en) * | 2012-03-06 | 2012-08-01 | 中国科学院工程热物理研究所 | Fan optimization arrangement method based on binary particle swarm optimization (BPSO) |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2550452B1 (en) * | 2010-03-23 | 2014-06-11 | Vestas Wind Systems A/S | A method for de-icing the blades of a wind turbine and a wind turbine with a de-icing system |
-
2013
- 2013-11-23 CN CN201310603969.3A patent/CN103603776B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101886617A (en) * | 2010-06-07 | 2010-11-17 | 三一电气有限责任公司 | Wind generating set and blade deicing system thereof |
| CN102003353A (en) * | 2010-12-10 | 2011-04-06 | 重庆大学 | Deicing method for blades of large-scale wind driven generator |
| CN102622482A (en) * | 2012-03-06 | 2012-08-01 | 中国科学院工程热物理研究所 | Fan optimization arrangement method based on binary particle swarm optimization (BPSO) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019086287A1 (en) * | 2017-10-30 | 2019-05-09 | fos4X GmbH | Method for forecasting the yield of a wind farm under icing conditions |
| DE102017125457B4 (en) | 2017-10-30 | 2023-02-23 | fos4X GmbH | Method for determining a probability of throttling and/or shutting down at least one wind turbine due to ice build-up |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103603776A (en) | 2014-02-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Gao et al. | A field study of ice accretion and its effects on the power production of utility-scale wind turbines | |
| CN103603776B (en) | A kind of prediction algorithm of wind generating set pitch control Security | |
| CN102410140B (en) | Ice load operation optimization control system and method used in wind turbine generator system (WTGS) | |
| CN102817780B (en) | Device and method for controlling freezing of wind generating set | |
| CN104454386B (en) | Icing control method and device for wind generating set | |
| WO2012119061A1 (en) | Methods for predicting the formation of wind turbine blade ice | |
| CN101737245A (en) | Blade pitch management method and system | |
| CN103216383B (en) | Control system and control method for wind generating set | |
| CN102434392A (en) | Wind speed and wind direction sharing system of wind electric field set | |
| CN202280571U (en) | Ice-load operation optimizing control system of wind generating set | |
| CN103590972B (en) | Variable pitch self-protection method for wind turbine generator set | |
| CN113007041A (en) | Wind turbine generator blade icing detection system and detection method | |
| Ahmed | Novel developments towards efficient and cost effective wind energy generation and utilization for sustainable environment | |
| CN108386316B (en) | Wind power generating set noise reduction system and method, control device and its storage medium | |
| CN203161453U (en) | Device used for detecting whether wind turbine generator blades are frozen | |
| Corradini et al. | A robust observer for detection and estimation of icing in wind turbines | |
| CN203452981U (en) | Draught fan with blade hot gas anti-congelation system | |
| CN107795440B (en) | Blade icing alarm method and device | |
| CN108425812A (en) | A kind of wind generator set blade icing flexible measurement method | |
| CN106677985A (en) | Wind turbine generator set evaluation system and predictive control service system thereof | |
| CN202768252U (en) | Freezing control device of wind turbine generator system | |
| CN115306655A (en) | Blade Internet of things wireless monitoring device and early warning method | |
| Jankoo et al. | Fuzzy logic control mechanism for flash flood monitoring station | |
| Sakipova et al. | Sail-Type Wind Turbine for Autonomous Power Supplay: Possible Use in Latvia | |
| Yan et al. | Numerical simulation of icing effects on static flow field around blade airfoil for vertical axis wind turbine |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20190614 Address after: 116601 Room 1505, No. 9-1 Haifu Road, Dalian Free Trade Zone, Liaoning Province Patentee after: Xiying Science and Technology (Dalian) Co., Ltd. Address before: 116600 Dalian Shanneng Science and Technology Development Co., Ltd. of No. 10 ShuangD Fifth Street, Dalian Development Zone, Liaoning Province Patentee before: Dalian Shinergy Science and Technology Development Co., Ltd. |
|
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20190731 Address after: 116600 Dalian City, Liaoning Province, China (Liaoning) Free Trade Pilot Zone Dalian Economic and Technological Development Zone Life Road 39-13-2, -3, -4 Patentee after: Dalian Shangjia New Energy Technology Co., Ltd. Address before: 116601 Room 1505, No. 9-1 Haifu Road, Dalian Free Trade Zone, Liaoning Province Patentee before: Xiying Science and Technology (Dalian) Co., Ltd. |